We are experimenting with the Hgmm Point Cloud Matching. we are noticing that it has a lot of drift on the xyz output in websight. How do we correct this?
We also notice that it creates a pose in the pose three that is hgmmTodom, and also produces a delta xyz. My assumption is that this matcher is somehow using odometry (but it is not documented) to create the delta. is the purpose of this to create a create a better pose estimation fusing odom and the point cloud difference together?
Can I have a better explanation of what is actually going on here and how to get a usefull pose output from this.